More accurate charge temp estimation (like Autronic)

Any feature requests for the Syvecs Engine Control Systems
EMS
Posts: 41
Joined: Tue Oct 12, 2010 2:24 pm

More accurate charge temp estimation (like Autronic)

Post by EMS »

I would like to see a table like Autronic, for more accurate estimation of charge temp. When having low charge velocity, the effective charge temp is close to ECT in stead of ACT. After mapping the fueling nicely to the desired value at low load/RPM and leave the car idling for a bit of time and let the ACT rise, the mixture gets leaner, this is because the ECU corrects for rising ACT. But actual charge temp is hardly changing at idle, because of rising ACT...

Autronic has solved it nicely (in my opinion) with a table which gives the "correct" charge temp estimate. I have copied the Autronic Manual to make it better to understand:

"Charge temp estimate" is the contribution of coolant & air temperature to make up "Charge temp". If this table is correct, air and coolant temperature changes (With in normal warmed up engine running conditions), will not effect the A/F ratio.

Contribution example.
Charge temp estimat=0%(Charge temp=Air temp)
Charge temp estimat=100%(Charge temp=Coolant)

Higher percentage makes "Charge temp" more bias to coolant temp. The higher the load & rpm, the smaller the "Charge estimate" value should be. This is because the air has less time to be warmed by the coolant.

NOTE1
Water cooled engines, vary the air temp & trim the "Charge est" so that the mixture does not change.

Air cooled engines, vary the engine temp (After a little cool down).

NOTE2
"Water cooled engines (Constant water temp)".
Higher air temp=Leaner mixture then INCREASE "Charge est".

"Air cooled engines (Constant air temperature)".
Higher engine temp=Leaner mixture then DECREASE "Charge est"

NOTE3
SM4-Default
LOAD(%)\Engine speed(Rpm)=Charge temp estimate(%)
0 1000 4000 7000
0.0 99.6 60 50 40
100.0 99.6 50 40 30
200.0 99.6 40 30 20
Benkku
Posts: 50
Joined: Tue May 03, 2011 12:30 am

Re: More accurate charge temp estimation (like Autronic)

Post by Benkku »

There is table (and possibility to use temp sensor for the purpose) for charge air temperature, which - in my opinion, covers this perfectly well.

Under load, charge air temperature is not related to charge air velocity at all, but pressure vs. temperature - which follows ideal gas law and compressor efficiency perfectly well.

'Charge temp' is - in my opinion somewhat 'misleading' term, there is internal engine friction, accessory loads and then there is actual VE - which is closely related to air temp. Air temp correction is more descriptive, volumetric content of oxygen mass, varies about 3.2 % per 10 degrees of celsius temperature change - altougth not linearly.

Where - in your application, air temp related sensors are positioned?
EMS
Posts: 41
Joined: Tue Oct 12, 2010 2:24 pm

Re: More accurate charge temp estimation (like Autronic)

Post by EMS »

There is a table (of course) for AIT correction on fueling in the Syvecs. But as I have written above, a table like the Autronic would help at low load/RPM to have more stable AFR. At full load there is no need for such table.

Mark.
Benkku
Posts: 50
Joined: Tue May 03, 2011 12:30 am

Re: More accurate charge temp estimation (like Autronic)

Post by Benkku »

Neither at low load, that's my experience with this ecu. AFR/lambda stays nicely at defined value. At least, i had no difficulties mapping despite of largely varying air temp (-20 deg c to 55 deg c), ecu follows nicely.

You wrote...
"After mapping the fueling nicely to the desired value at low load/RPM and leave the car idling for a bit of time and let the ACT rise, the mixture gets leaner, this is because the ECU corrects for rising ACT. But actual charge temp is hardly changing at idle, because of rising ACT..."

Could you explain - what your air temp sensor then measures, and what is the reasoning behind, that air temp correction table do 'not work' (you said leaning out). Or is the temp sensor thermal mass 'big enough', not to react to fast changing air temp? Think about those... In my opinion - air temp sensor is positioned wrongly, if it measures 'something else', than actual inlet air temp. If it does not react - well, then sensor thermal mass is too big for the application.
EMS
Posts: 41
Joined: Tue Oct 12, 2010 2:24 pm

Re: More accurate charge temp estimation (like Autronic)

Post by EMS »

My ACT sensor is placed in the throttle body, even at very small throttle opening the air speed over the sensor is high and it measures perfectly the temperature of the air which enters the inlet manifold. This is as far "downstream" I can measure ACT properly.

But I think you don't understand what I mean... At low RPM/load, there is a MASSIVE difference between the ACT measured just before the inlet manifold and the actual temperature of the air entering the cilinders! This is where the "problem" is. The Syvecs ECU only corrects for ACT measured just before the inlet manifold (or inside the inlet manifold). Which is absolutely fine for higher load/RPM, but when airflow is very low, there is a big difference between REAL ACT and measured ACT which can cause AFR to change.
Benkku
Posts: 50
Joined: Tue May 03, 2011 12:30 am

Re: More accurate charge temp estimation (like Autronic)

Post by Benkku »

Yes i know 'that phenomen' from some daily drivers, where inlet manifold temp raises through convection. Consider repositioning air temp sensor, to the plenum near cylinder runner/inlet track opening. Inlet manifold thermal barrier coating/painting and plating (ie. polymer plate under inlet gasket) also helps much. Aluminum has rather high thermal conductance, which in technical plastics typically has over 500 times lower.

Al thermal conductance 237 (W/m-K)c

some plastics...
PEEK k = 0.25 (W/m-K)c (luja tekninen muovi, glass temperature 143 deg. C)
PEEK GF30 k = 0.43 (W/m-K)c (luja tekninen muovi, glass temperature 143 deg. C)
PC GF30 k = 0.24 (W/m-K)c (luja tekninen muovi, glass temperature 150 deg. C)
Polypropylene (PP) k = 0.12 (W/m-K)c
Polyisoprene k = 0.14 (W/m-K)c
Polymethyl methacrylate (PMMA) k = 0.167-0.25 (W/m-K)c
Bakelite k = 0.15 (W/m-K)c
Teflon k = 0.25 (W/m-K)c
Nylon 6,6 k = 0.24 (W/m-K)c
EMS
Posts: 41
Joined: Tue Oct 12, 2010 2:24 pm

Re: More accurate charge temp estimation (like Autronic)

Post by EMS »

Moving the AIT sensor close to the inlet port in the cylinder head will help, but the Autronic solution works really well even with the sensor placed a bit further from the head(s).
littlefella
Posts: 14
Joined: Wed Nov 09, 2011 9:11 am

Re: More accurate charge temp estimation (like Autronic)

Post by littlefella »

Autronic charge temp ESTIMATION.. What a total load of crock.. ESTIMATION.. when and how do you ESTIMATE the difference in charge temp from where do you "calculate" that.. Do you EATIMATE the ignition advance?? I have used many ecu's including Autronics and NEVER found any need for this. WIth Autronic I always cancel this totaly stupid function, it can lead to problems with big boost engines, or high temperature changes, forests into hot open spaces etc.. I dont comment on things much, but this one I will.. I have Syvecs ecu's running here in Czech at -15c and the lambda (without closed loop) is the same as I set it when I tuned the engine at 24c! 25years of experiance and mapping in controled cells, and in Artic conditions and Deserts, from Dead sea to 5000m in Peru gets you the correct data, not ESTIMATIONS.
Cheers
EMS
Posts: 41
Joined: Tue Oct 12, 2010 2:24 pm

Re: More accurate charge temp estimation (like Autronic)

Post by EMS »

It works PERFECTLY well if you set it up right!!! I don't see why it should give a problem with high boost engines, as from anything above idle and slowly driving with low airflow (up to around 800 Mbar absolute and around 4000 RPM) there shouldn't be any difference with the Syvecs. (when set up properly that is) It is quite a bit of work to setup, but when done once, it can be used for similar cars.

If you don't like Autronic ECU's, don't use my topic to show that... ;)
Benkku
Posts: 50
Joined: Tue May 03, 2011 12:30 am

Re: More accurate charge temp estimation (like Autronic)

Post by Benkku »

Proceed with those changes i recommended - of course, the air temp table needs to be correctly defined.

You said that engine is 'leaning out' - which means, that actual engine drawn air temperature, is lower than measured and/or air temp correction table is wrongly defined.
Post Reply